pipeline_tag: text-generation language: - en datasets: - allenai/c4
This is a 800M parameters model pre-trained with QuEST over 80B C4 tokens in INT4 format (W4A4).
The code to verify that this model works in INT4 can be found here.